AI is now creating poems, images, music, and designs in ways that look increasingly “human.” That shift creates a central copyright problem: if an AI produces the output, who, if anyone, is the author?
Around the world, courts and lawmakers are trying to fit generative AI into copyright rules that were built on a human-authorship model. The US approach is strict: copyright requires a human author, so AI can’t be credited as the creator. Other places are exploring middle-ground models that protect works where a human meaningfully guides or shapes the AI’s output.
India is moving from abstract debate to real litigation. Under the Copyright Act, 1957, authorship is tied to natural persons, so AI—lacking legal personality—cannot be an “author” or hold rights. That means purely machine-generated works are unlikely to qualify for copyright protection unless the law changes.
However, India may still protect AI-assisted works if a human plays a substantial creative role—by directing the process, exercising creative judgment, and editing or curating the final result. In that scenario, the human can be treated as the author because the output reflects human intellectual effort rather than the autonomous creation of a machine.
Different jurisdictions illustrate different policy choices:
- United States: insists on human authorship; AI-only images or text aren’t protected, though a larger work may be protected if it includes sufficient human-created elements.
- United Kingdom: provides that for computer-generated works, the author is the person who made the necessary arrangements for the work’s creation.
- European Union: has been considering a “significant human input” approach—protecting work where humans meaningfully influence the outcome.
- China has sometimes granted protection where originality can be linked to identifiable human input (such as careful prompting, selection, and modification).
Beyond ownership, AI raises broader copyright pressures: whether training on copyrighted material is lawful (and under what exceptions), how attribution and moral rights work when a machine is involved, and how businesses can safely commercialise AI outputs amid legal uncertainty. Some propose that AI-generated works should default to the public domain; others argue for a new, tailored (“sui generis”) right for AI content.
For India, potential reform options include defining a clear “significant human involvement” test, establishing licensing regimes for training data to ensure creators are compensated, consulting widely with industry and artists, and aligning with international standards to protect Indian creators in global markets.
The piece highlights two illustrative disputes: ANI v OpenAI (2024) in India, framed as a test case about training on copyrighted journalism and alleged reproduction; and GEMA v OpenAI, framed as a European test about whether models that can reproduce protected lyrics have infringed copyright. Overall, the theme is the same: law must balance innovation with the protection of creators’ incentives and rights.
Summary (key points)
- Core issue: Copyright law is built around human authorship, but generative AI blurs that line.
- Main legal questions:
- Who owns AI outputs: user, developer, nobody, or a shared model?
- Is training on copyrighted works infringement or covered by exceptions?
- Should machine-generated outputs be unprotected/public domain, or get new tailored rights?
- India’s current position (as described): AI can’t be an “author”; protection generally requires human creativity under the Copyright Act, 1957.
- Practical workaround: AI-assisted works may be protected if a human meaningfully shapes, edits, or curates the output (human-as-author via substantial input).
- Global approaches vary: the US is strict on human authorship; the UK assigns authorship to the person who makes the necessary arrangements; the EU discusses “significant human input”; and China has, in some cases, treated AI as a tool and protected outputs with demonstrable human originality.
- Policy direction suggested: India may need clearer tests for human involvement, licensing frameworks for training data, stakeholder consultation, and international alignment.
- Big picture: The “right” solution is a societal choice about how to reward creativity while enabling AI innovation.
source: Depenning & Depenning
Leave a Reply